Estimating the Abundance of South Baffin Caribou

Summary Report 2012

Deborah A. Jenkins1

Jaylene Goorts2

Nicolas Lecomte3

1 Regional Wildlife Biologist, Department of Environment, Government of Nunavut PO Box 400, Pond Inlet, Nunavut, X0A 0S0

2 Wildlife Technician, Department of Environment, Government of Nunavut PO Box 400, Pond Inlet, Nunavut, X0A 0S0

3 Ecosystem Biologist, Department of Environment, Government of Nunavut PO Box XXX, Igloolik, Nunavut, X0A 0L0

i

TAKE-HOME MESSAGE

Based on the results of a large-scale aerial survey, conducted from March to May 2012, we

estimate that the abundance of caribou on south Baffin Island was between 1065 – 2067 (95%

CI) animals aged one year or older. Across the study area, we tallied 185 observations of caribou

groups both on and off transect; representing a total of 518 individuals. Inuit knowledge collected

in 10 communities also identified low numbers of caribou on Baffin Island and together these

results highlight the need for a comprehensive co-management system and conservation plan.

SUMMARY

Caribou are a critical component of the boreal and arctic ecosystems; they are culturally significant to

local communities and provide an important source of food. At a circumpolar scale, there is a general

decline in caribou population and this could have major consequences on the traditional harvesting

that occurred for millennia, especially for Inuit. In some areas, there is still uncertainty on population

trends because of the lack of scientific information due to logistics and remoteness; this is

particularly true for Baffin Island, where three populations of Barrenground caribou are currently

recognized and little is known about their abundance and trends over time (e.g. Ferguson and

Gauthier 1992). In the past 60 years, only discrete portions of their range have been surveyed and no

robust quantitative estimates at the population level were ever derived. For over a decade Inuit from

communities on northern Baffin Island, and more recently from across the entire island, have

reported declines in caribou numbers on Baffin Island, although no quantitative estimates are

available. Here we report results of a caribou survey conducted from March to May 2012 across the

range of South Baffin caribou; the largest range occupied by one of the three caribou populations of

Baffin Island, Nunavut. The objective of the survey was to provide the first comprehensive

population estimate and update current knowledge on their status. We used a distance sampling

framework with line transects flown by two helicopters. Survey lines were spaced 10 km apart (to

account for the large study area), which were then stratified by ecoregion, and oriented east–west

across the study area. Following Buckland et al. 2001, spacing was chosen to balance precision,

sufficient line length, and survey resources. Overall, we flew over 27,000 km of transect and saw 143

groups of caribou for an encounter rate of approximately 5 groups per 1000 km of flying. Using

Distance 6, we estimated that the South Baffin region hosted between 1,065 - 2,067 (95% CI))

ii

caribou one year or older in March-May 2012. To be conservative, the analysis did not include short-

yearlings which would have increased this estimate. As well, the survey was completed before

calving and therefore did not estimate the productivity of the population in 2012. Previous

projections of population size range between 60,000 and 180,000 for early 1990’s (Ferguson and

Gauthier 1992, see also Jenkins and Goorts 2013, Jenkins et al. 2012); thus, our results represent a

>95% decline in abundance in a 20-year window. Given the critical impact of caribou on local

communities and the Arctic ecosystem at large, this situation emphasizes the requirement for a

comprehensive co-management system and conservation plan.

iii

ACKNOWLEDGEMENTS

Funding for this study was provided by the Department of Environment, Government of Nunavut,

Polar Continental Shelf Project, and the Nunavut Wildlife Management Board. In-kind support was

provided by the Canadian Wildlife Service, Baffinland Iron Mines and Peregrine Diamonds. Thanks

to field personnel, including Chris Wex, Ben Widdowson, Paul Idlout, Oquituk Ashoona, Jaypootie

Akpalialuk, Elishua Kango, Milissa Elliott, Guillaume Szor, Kristi Lowe, Jean Daniel Blouin, Nick

Burnaby, Sara Holzman, Patricia Peyton, and specially Mitch Campbell for his crucial role in the

survey. To our helicopter pilots, Gerard Hartery, Daryl Hefler, Jimmy Abbott, and John Innis and

mechanical engineers, Rainer Kemming, Chris Organ, Jason Simms, Garland Pope, and Mark Foley,

all of Universal Helicopters, a big thank you. A special thanks to Pat Rakowski for managing the

remote field camp and providing outstanding meals and hospitality, and Jim Leafloor, CWS, for his

program-saving support. For logistical support our gratitude to Jimmy Noble, Peter Hale, Alden

Williams, Guillaume Szor, Theo Ikummaq, Chris Wex, Jeffrey Qaunaq, Sam Nuqingaq, Tim Cater,

Malik Awan, Nicolas Lecomte, Sharina Kennedy, Mike Mifflin, Jerry Arreak, Mathewsie Onalik,

Amy Aglak, Tyler Ross, Kristi Lowe, Willy Flaherty, Steve Atkinson, and especially Milissa Elliott.

A special thanks to Tim McCagherty, Glenn Parsons, and Jodi MacGregor of PCSP for daily

logistical support. Review comments provided by Tiago Marques and Eric Rexstad, Centre for

Research into Ecological and Environmental Modelling, University of St. Andrews, Mathieu

Dumond, Department of Environment, Government of Nunavut and Seth Stapleton, all contributed to

improve this report.

iv

TABLE OF CONTENTS

1.0 BACKGROUND…………………………………………………………………………….......7

2.0 MATERIALS AND METHODS.................................................................................................8 2.1 Study Area……………………………………………………………………………….8 2.2 Survey Design……………………………………………………………...…………….9 2.3 Field Methods…………………………………………………………………….…….10 2.4 Analysis…...…………………………………………………………………………….11

3.0 RESULTS……………………………………………………………………………………….13

3.1 Observations……………………………………………………………………………14 3.2 Exploratory Analysis……………………………………………………..……………14 3.3 CDS and MCDS Detection Function Models…………………………...……………14

4.0 DISCUSSION………………………………………………………………………….………..15

4.1 Population Estimate………………………………………………………..…………..15 4.2 Evaluation of Methodological Assumptions…………………………………….……16 4.3 Consequences……………………………………………………………………...……18

5.0 REFERENCES……………………………………………………………………………..…..20

6.0 TABLES AND FIGURES………………………………………………………………...……24

v

LIST OF FIGURES Figure 1a. Delineation of Barren-ground caribou populations on Baffin Island, as currently recognized (modified from DOE 2005)………………………………………………………………25 Figure 1b. South Baffin Island caribou study area, 2012……………………………………………25 Figure 2. Ecoregions (as described by the Ecological Stratification Working Group, 1995) within the South Baffin study area (2012)………………………………………………………………….……26 Figure 3a. South Baffin survey transect lines, spaced 10km apart and stratified by ecoregion.…….28 Figure 3b. Combined survey tracklogs, recorded from March 27th to May 27th, 2012 by GPS units mounted in each helicopter. Three islands in Hudson Strait and 4 islands in Foxe Basin were not completed due to open water and inclement weather……………………………………………...…28 Figure 4. Detection probability (continuous line) plot and histogram of perpendicular distances from the transect line for clusters of caribou in the South Baffin survey area, March 27-May27, 2012. The g(x) is estimated using a Half Normal model (see Table 5, below). Bin size is 165 m……………...31

vi

LIST OF TABLES Table 1. Area (sq. m) surveyed in each ecoregion in the South Baffin survey area…………………27 Table 2: Overall tally of observations and tracks recorded on and off- transect………………….…29 Table 3. Composition of caribou seen On and Off transect……………………………...……….…30 Table 4. Details of the MCDS and CDS Models considered for the detection function, with available covariates……………………………………………………………………..………………………32 Table 5. Estimates of caribou density and abundance within areas of eight different strata cover by line transect survey, based on analysis in the program Distance 6……………….…………………..33

7

1.0 BACKGROUND

Barrenground caribou on Baffin Island are a highly valued food source and subject to harvesting

by 10 communities, including Iqaluit, the capital of Nunavut. Since the mid-1990s hunters and

local Hunting and Trapping Organizations (HTO) have regularly reported that caribou were

increasingly difficult to locate and that numbers were low (Jenkins and Goorts 2013, Jenkins et

al. 2012). Barrenground caribou are the only ungulate that inhabits Baffin Island and three

populations are currently recognized: South Baffin, North Baffin, and Northeast Baffin ( DoE

2005, Ferguson and Gauthier 1992, Ferguson 1989; Figure 1a). These caribou are unique

compared to other Barrenground herds, as they do not overwinter in forested habitat, nor do all

caribou undertake long seasonal migrations to calving areas (Ferguson 1989, 1998, 2003,

Jenkins and Goorts 2011, Jenkins et al. in prep). Thus, Baffin caribou have been defined as

populations and not as herds, with initial delineations based on areas where animals

concentrated in the fall to breed (Ferguson 1989). In addition to Inuit Knowledge, tagging

studies (from 1974 - 1984), and extensive caribou locations from collaring research (1987-1994

and 2008-2011), provided scientific evidence to support the separation in North and South

Baffin populations (Jenkins et al. in prep., Jenkins and Goorts 2011, Ferguson et al. 1998,

Ferguson 1989, Kraft 1984). The Northeastern population refers to caribou occupying the

headwaters and fiords along the northeastern coast, and is based mainly on Inuit Knowledge

(Ferguson 1989). The South Baffin caribou range extends across more than half the Island and

supports both sedentary and migratory animals (Ferguson 1989, 1998, 2000, 2003, Redhead and

Land 1979, Elliot and Elliot 1975, Tener 1960). Thus, population refers to groups of individuals

that occur within a geographic range, leading to distinct demographic characteristics. Evaluation

of population structure at the genetic level is underway to better understand these delineations

(e.g. into three populations) and how they are possibly linked to each other.

Climate change, disease, and anthropogenic activities, including mineral development

and unrestricted harvest, have been identified as possible threats to South Baffin caribou

(Jenkins et al. 2012, Festa-Bianchet et al. 2011, Jenkins and Goorts 2011, Jenkins 2011a, Rippin

1972, Tener 1960, Kelsall 1949, Wright 1944). In the same context, Joannasie Kooneeliusie, an

Elder from Qikiqtarjuaq, has indicated that climate change is playing an increasingly significant

role in the decline of caribou (Jenkins et al. 2012). Yet, no quantitative estimate of caribou

abundance, based on any study that covers the entire South Baffin range, has been available until

8

now. Published values for all Baffin Island caribou include Williams and Heard (1986) who

reported that > 100,000 caribou inhabited the Island in 1985. The status was updated for 1991

when Ferguson and Gauthier (1992) proposed that caribou populations were stable with 60,000-

180,000 caribou in South Baffin, 50,000-160,000 caribou in North Baffin, and >10,000 caribou

in the Northeast (Figure 1a). However these estimates are not based on any systematic survey;

they were derived from incidental observations, local knowledge, anecdotal information, and

professional opinion (Ferguson and Gauthier 1992).

One of the most widely used methods for estimating both the abundance and density of

wildlife is Distance Sampling (e.g. Aars et al. 2009, Buckland et al. 2004). This method is

particularly suited to populations occurring at low densities over large geographical areas

(Buckland et al. 2004). For these reasons, Jenkins (2009a, 2010a, 2011a, 2012) advocated a

Distance Sampling, line-transect aerial survey for Baffin Island caribou, highlighting as benefits:

1) the ability to utilize unbounded strip widths, 2) the method-based assumption that not all

animals are seen, and 3) the ability to produce robust density estimate after a single survey. In

the present paper, we report on the first year of a large-scale, multi-year line transect survey of

caribou on Baffin Island, Nunavut (see Jenkins 2011a). The survey results are for South Baffin

caribou and the period March-May 2012.

2.0 MATERIALS AND METHODS

2.1 Study Area—.Baffin Island forms the eastern margin of the Canadian Arctic Archipelago

and is the largest island in Canada (ca. 507,451 km2). This island represents the northern fringe

of Barrenground caribou range and caribou generally occur in small groups scattered over large

geographical areas (Jenkins and Goorts 2011, Jenkins 2011, Jenkins 2010, Elliot and Elliot

1975, Tener 1960).

The 2012 study area encompassed the range of South Baffin caribou; the largest range

occupied by one of the three caribou populations of Baffin Island (Figure 1a). The population

range was previously delineated by Ferguson and Gauthier (1992) and the Government of

Nunavut (DoE, 2005), and we used these boundaries, with slight adjustments based on caribou

location data (1984-2011; Jenkins 2011a, Jenkins and Goorts, 2011, Jenkins et al. in prep) and

9

several community consultations (Jenkins et al. 2012), to define the 2012 study area (Figure 1b;

see Jenkins 2011). For instance, Daniel Qattalik, from Igloolik said “When he was a child they

would go caribou hunting just north of the 2012 study line, and they would go south of study

line. He thinks that the study line is bang on and that caribou from the north generally do not

cross down into the south and vice versa with caribou in the south”.

The area, including southern Baffin Island and proximal islands in Foxe Basin, Hudson

Strait, and Davis Strait, was stratified into 9 ecoregions based on bioclimatic characteristics as

defined by the Ecological Stratification Working Group (1995; Figure 2). The area of each

ecoregion ranged from 2,452 km2 (Wager Bay Plateau) to 71,844km2 (Meta Incognita

Peninsula) and totalled 289,715 km2 (Table 1). Two small islands, historically considered part of

North Baffin caribou range, were included in the 2012 study area, to explore possible use by

South Baffin caribou.

There were 4 communities within the study area, which are Pangnirtung, Iqaluit,

Kimmirut, and Cape Dorset. The area also hosts a National Park (Auyuittuq National Park) and

a Migratory Bird Sanctuary (Dewey Sober) (Figure 1b).

2.2 Survey Design --.We used a combination of historical survey data, harvest tag data, radio

collar location data, Inuit knowledge, and consultation to inform survey design (Jenkins and

Goorts 2013, Jenkins et al. 2012, Jenkins et al. 2011, Jenkins 2010a, Ferguson et al. 1998, Kraft

1984, Chowns 1979, Redhead and Land 1979, Freeman, M.R. 1976, Elliott and Elliott 1974,

1975, Jenkins et al. in prep). Caribou were known to occupy Baffin Island, where glaciers,

mountains, and steep river valleys appear to influence the geographical separation of south

Baffin caribou from the north and northeast (Jenkins et al. in prep, Jenkins and Goorts 2013,

Ferguson and Gauthier 1992, DoE 2005), but also proximal island in Foxe Basin, Davis Strait,

and possibly Hudson Strait (Jenkins et al. 2011, Jenkins 2010a, Ferguson et al. 1998, Elliott and

Elliott, 1974, 1975). Using Arc Map 10.0 (www.ESRI.com), we developed a systematic,

stratified line transect design with a random starting location. Lines were stratified by ecoregion,

spaced 10 km apart, and oriented east-west across the study area (Figure 3a). Following

Buckland et al. 2001, spacing was chosen to balance precision, sufficient line length, and survey

resources. The orientation of lines was selected to place them perpendicular to any potential

density gradients and known movement patterns (Buckland et al. 2011, DoE Unpublished data,

10

Jenkins et al. in prep). From ArcMap, the line transect coordinates were uploaded into GPS units

(Garmin GPSMAP 196, 276C) used in the helicopters during the survey. The transect lines

covered the entire land base with the exception of extensive ice fields or glaciers (Figure 3a) and

were flown using a conventional single platform approach (Buckland et al. 2001).

The survey was scheduled for late winter (mid-March to mid-April) during the pre-

calving period, when caribou movement is low (average 2.14 km/day, SE 0.14; estimate based

on South Baffin caribou collaring data, March 15-April 15, 1987-1991; Jenkins et al. in prep),

light conditions and temperatures permit safe operation of helicopters, and snow provides a good

backdrop for sighting wildlife and their sign.

2.3 Field Methods--.The systematic line transect survey was flown using 2 Bell 206 LR rotary

wing aircraft at approximately 120 meters above ground level and an average speed of ca. 140

km/h to maximize detectability. Helicopters were initially stationed together at Iqaluit

(63.729943 N, -68.448588 W), and then separately, at the communities of Pangnirtung

(66.144279 N, -65.71706 W) and Cape Dorset (64.232268 N, -76.542675 W). By mid-May, the

helicopter teams reunited at Nikko Island Research Camp (66.596447 N, -71.526536 W) (Figure

1b). One helicopter was released on May 17th while the other repositioned to Mary River Camp

(71.3264 N, -79.3762 W), to complete the most northerly portion of the survey area. Each

helicopter was staffed by 4 observers, including the pilot: the two forward observers (pilot and

caribou biologist) focused on the transect line in front of the helicopter, while the two rear

observers (wildlife technician and local HTO representative) focused to each side of the

helicopter with some overlap with the front observers search area. The helicopter teams were

reduced to three observers during a staff change-over (e.g. 4 days) and when transitioning

between base camps (i.e. 2 part-days). During this time, one observer focused forward, while the

two remaining observers focused to each side of the helicopter, with increased overlap with the

forward observer. We do not think that this affected our ability to locate caribou as this extra

effort was set to compensate for the lack of one observer. We used an unbounded line transect

design where transect width was infinite; therefore, all caribou observations where collected,

regardless of distance from the transect (Buckland et al. 2001).

Caribou observed while flying along the transect line were considered on-transect.

Observations collected while ferrying to or between transects, to wildlife sightings, or to fuel

11

caches where considered off-transect. Because caribou are gregarious and can occur in well-

defined clusters, each observation represented a group of caribou, defined here as one or more

animals located within ca. 100 m of each other (following Jenkins et al. 2011). We investigated

all observations by moving off the transect line to record the location (e.g. at the center of the

group where they were first observed), group size, and sex/age classes. Caribou were classified

as male or female, and as adult (>2 years), yearling (>1 year and < 2) or calf (<1 year). Animal

care and safety was a priority and observation time was kept to a minimum to reduce

disturbance. For each observation, habitat structure was recorded as a covariate and scored at a

coarse and fine spatial scale. At the coarse scale (beyond 100 m radius of the observation), the

terrain was scored as F=Flat, R=Rolling, or M=Mountainous. At a finer scale (within 100 m

radius of the observation), the terrain was scored as 1=Flat, 2= Moderately Sloped, or 3=Steep.

All spatial data were collected on hand held GPS units (Garmin GPSMAP 196, 276C, 76S, and

78C, Oregon 250T, and 450T), which also recorded helicopter positional information every 20-

30 seconds to produce detailed track logs for each flight (Figure 3b). When animal care,

environmental conditions, fuel and time permitted, scat samples of caribou were collected for

genetic investigations.

To support distance sampling analysis, our survey platform, survey design, and

geoprocessing operations, were designed and executed to meet the three key assumptions of

distance sampling (below), and produce an unbiased estimate of density (Anderson et al. 2001;

Buckland et al., 2001); 1) all animals of interest that were directly on the transect line were

detected; 2) animals of interest were detected at their initial location before they moved in

response to the observer (i.e. away from the aircraft); and 3) perpendicular distance (y) from the

transect line to each detected cluster was measured accurately. In the discussion, we provide an

assessment of the degree to which each of these was accomplished.

2.4 Analysis--.Survey data were integrated in a Geographical Information System and we used

Arc Map 10.0 to measure the perpendicular distance (y) from each wildlife observation to the

transect (represented by the actual flight line: Marques et al. 2006). The length of the flight lines,

estimated by ecoregion, was also measured in Arc Map 10.0 (Aars et al. 2009, Jenkins et al.

2011). To minimize any measurement error, we used a North Pole Azimuthal Equidistance

Projection centered on each ecoregion and conducted all line measurements at a scale of

1:10,000 (precision +20 m); and distance measurements at a scale of 1:5,000 (precision +10 m)

12

We followed Buckland et al. (2001, 2004) and used Program Distance, Version 6.0

(Thomas et al. 2009), to model the line transect data and estimate density and abundance for

caribou. Using the standard or conventional distance sampling approach (CDS), we modeled the

probability of detecting a group of caribou, and their density, as a function of distance alone.

That is, the detection function represents the probability of detecting a group of caribou, given

that it is at distance (y) from the transect (Buckland et al. 2001). Recognizing that other

variables, may affect the detection probability, density estimates were also derived using

multiple covariate distance sampling (MCDS), which allowed us to model probability of

detection as a function of both distance and one or more additional covariates (Buckland et al.

2004). This approach was explored in order to increase the reliability of density estimates made

on subsets of the data (e.g. geographic stratum which are identified here by ecoregion), and to

increase precision of the density estimates (Marques et al. 2007). We used detection function

models (key function/series expansions) recommended by Marques et al. 2007 and Buckland et

al. (2001, 2004). Exploratory data analysis followed Buckland et al. (2001), and we right

truncated (w) at the distance 2,800 m in all analysis, based on recommendation to delete outliers,

address size bias in detected clusters and facilitate modeling of the data (two observations

excluded so this has little effect on the estimates). To estimate density and abundance by

geographic stratum, we used a global model for the detection function (as the number of

observations were too limited to fit a separate detection function for each stratum), and then

used observation level covariates to improve the model (as recommended by Marques et al.

2007, Buckland et al 2001, 2004).

We used a global cluster size that was estimated using a size bias regression method

when regression was significant at an alpha level of 0.15. If the regression was not significant,

the global mean cluster size of observations was used. We estimated variance using a

nonparametric bootstrap with 999 resamples generated by sampling with replacement from the

lines (Marques et al. 2007). We used Akaike’s information criterion (delta AIC), to select the

model with best fit. Such model was accepted if it had a non-significant goodness of fit value

(Chi Square), a non-significant Kolmogorov-Smirnov Test, and non-significant Cramer-von

Mises test (Buckland et al. 2001, 2004).

13

3.0 RESULTS

The survey was flown in ca. 377 hours (including ferry flights) between March 27 and May 27,

2012. The transect lines and actual flight lines are shown in Figures 3a and 3b, respectively. We

flew 27,250 km on transect, starting from the far south and moving north. The survey area

included the non-glaciated portions of the South Baffin caribou range on Baffin Island, and

proximal islands (Table 1 and Figure 3b). In Hudson Strait, Mill Island, Salisbury Island, and

Nottingham Island, all within the Wager Bay Plateau ecoregion, could not be reached because of

open water and poor weather conditions. Ohito Ashoona, HTO Director, Cape Dorset, indicated

that it was not known if caribou inhabited the islands, as they were difficult to access (Jenkins et

al. 2012). However, information from multi-year polar bear surveys, and helicopter over-flights

indicated that no caribou signs had been observed from 2009-2012 inclusively (DoE,

unpublished data).

In the Foxe Basin, Koch Island and Rowley Island could not be reached for the same

reasons. Based on community interviews, Rippin (1972) reported that caribou do not occur on

either island. Additionally, during polar bear surveys in 2009-2010, and 2012 no caribou or

caribou sign were observed on Koch or Rowley Islands (DoE, unpublished data). These small

areas were subsequently excluded from the analysis although including them would decrease the

overall density for Baffin Island because of the absence of caribou (see Table 1). Almost

complete coverage of the remaining study area was achieved with the exception of a few lines

that were cut short due to fog, sheer mountain conditions, and/or ice crystals (Figure 3b). The

survey schedule was extended as unforeseen weather events created delays in our starting date

(from March 16 to March 27), in survey effort (periods ranging from 1-5 days), and the

positioning of one helicopter to the Nikko Island base camp (12 days). Notably, one of two

helicopter survey teams was able to position to the site and was maintain progress on the survey

during this time. The field work continued until late May when caribou movement to calving

areas (located within the study area; Ferguson 1995) is known to occur (Jenkins et al. in prep,

Ferguson 1989, Redhead and Land 1979, Elliot and Elliot 1975, Elliot and Elliot 1974). To

address the potential for delays, execution of the survey was planned from south the north, i.e. in

the direction of migration (Ferguson 1989, Redhead and Land 1979, Elliot and Elliot 1975,

Elliot and Elliot 1974). This ensured that calving areas were surveyed towards the end of the

field program and that any migrating animals were counted. Finally, survey progress along the

14

migration corridor was faster (ca. 40-50 km/day) than the average daily movement rate of

caribou (average 2.26 km/day, SE 149 m; South Baffin caribou collaring data; May 1987-1991;

Jenkins et al. in prep).

3.1 Observations--. Across the study area, we tallied 185 observations of caribou groups both

on and off transect; representing a total of 518 individuals (average 2-3 individuals per group),

including all age classes. Tracks were recorded on 83 occasions and 4 feeding sites were noted

(Table 2). By design, no newborn calves were encountered as the survey was completed prior to

calving. Sampling methods generated 143 observations of 324 individual caribou (at least one

year or older) on-transect. Composition of classified individuals encompassed a total of 155

female and 94 male adults, 72 yearling caribou, as well as 34 calves (also known as short

yearlings). Details are presented in Table 3. The proportion of calves is 9.6% of those animals

seen on-transect or 22 calves per 100 cows. The ratio of adult males to females is 61:100.

3.2 Exploratory Analysis --.Preliminary analysis in Distance 6.0 indicated no data problems or

difficulties, with special consideration for both evasive movement by caribou and a detection

probability on the flight line of < 1. Neither of these phenomena was detectable in the data.

Exploratory analysis of covariates revealed that some habitat structure categories were under—

represented. Thus, covariates were condensed into binary categories. Specifically, for large-scale

structure: Flat (F) =0, and Hills (H) and Mountains (M) = 1, and for fine scale structure: Flat (1)

= 0, and Moderately Sloped (2) and Steep Slope (3) = 1. Because course and fine scale habitat

structure were highly correlated (Pearson’s r=0.96, Fisher’s exact test, p<0.0001), only one

covariate was considered in the models at any time.

3.3 CDS and MCDS Detection Function Models --. Including either coarse or fine scale

habitat structure in the models resulted in a large improvement in model fit (delta AIC) over

CDS models. Based on delta AIC, we selected the Half Normal model as the best model (Table

4, Figure 4). This model had a good fit of the data (goodness of fit tests; 𝜒2=17.7411,

p=0.12378, Dn=0.0665, p=0.5615, and W2=0.1762, 0.300<p<=0.400). The estimated effective

strip width was 1,036 m (95% CI 886 - 1,211 m). The expected cluster size (or group size) was

2.13 (95% CI 1.91 - 2.38) whereas mean cluster size was 2.26 (95% CI 2.02 - 2.51) caribou per

cluster. We estimated the probability of detecting a cluster of caribou within the defined area as

15

Pa=0.37 (95% CI 0.32 - 0.43). The estimated density of caribou was 0.0053 caribou/km2 or 5.3

caribou per 1,000 km2 (95% CI 3.8 - 7.4/1000 km2). Based on our findings in the survey area

(South Baffin, 279,233.701 km2), we calculated an estimate of 1,484 caribou age one year or

older (95% CI 1,065 - 2,067; CV 0.17). To be conservative, the analysis did not include short-

yearlings which would have increased this estimate. As well, the survey was completed pre-

calving and therefore did not estimate the productivity of the population in 2012. Estimates, by

ecoregion, are provided in Table 5.

4.0 DISCUSSION

4.1 Population Estimate --.Overall, we flew over 27,000 km of transect and saw 143 groups of

caribou (ca. 5 groups per 1,000 km flown). For March-May 2012, we estimated with 95%

confidence intervals, between 1,065 and 2,067 caribou (one year or older) in the South Baffin

caribou range using Distance Sampling line-transect methods. The low number of observations,

dispersed over the large spatial area highlights the status of this caribou population, with an

estimated density that is only marginally larger than Peary caribou on Bathurst Island Complex

(0.0095 caribou/km2; 2001) and North and Southern Ellesmere Island (0.0083 caribou/km2;

2006 and 0.0092 caribou/km2; 2005, respectively; Jenkins et al. 2011). We found that caribou in

South Baffin occurred in small groups (average cluster size of 2, range in cluster size 1-16) and

this is consistent with other caribou studies within the Baffin Region, during late winter (see

Jenkins et al. 2011, Jenkins and Goorts 2011, Jenkins 2010, 2009). Given the extensive area

covered, tight transect spacing and sampling effort, it is reasonable to assume that a

representative sample of caribou was surveyed. Complete census of this huge area is not

possible and therefore, the distance sampling design was performed to capture the wide range of

these low-density caribou. Missing groups in between transect lines is inevitable (and assumed),

but the systematic design should promote an equal chance of detecting them.

Regarding the sex and age ratio values observed here, the interpretation is difficult as most of

the understanding of these parameters comes from fall surveys in other populations of caribou

and also because of the low numbers of caribou encountered on Baffin Island. Nevertheless, a

ratio of 22 calves per 100 cows, as measured in this study, is somewhat typical of population at

low-density and/or pertaining to a declining type (such as Peary caribou; e.g. Peary caribou

16

Status Report for the Northwest Territories, Species at Risk Committee). Also, the relative high

number of yearlings compared to adults, could be driven by higher harvest rate of adults and/or a

good breeding year. Community consultations will provide additional context for these

observations. .

4.2 Evaluation of Methodological Assumptions --.The survey design, including schedule was

carefully planned, to address the assumptions of Distance Sampling, produce an unbiased

estimate of density, and allow transects to be flown almost continuously during late winter (but

see below). Firstly, the design relied on flying a large number of randomly allocated line

transects within the survey area and these were successfully flown, with the exception of a few

short lines on small proximal islands. The small size of the islands and data from multiple

sources suggests that caribou were not present in these areas. Nonetheless, the area of the islands

was not included in the analysis, which would have lowered the density estimate if detection

was assumed as zero.

The survey schedule was extended as unforeseen weather events created delays in

survey progress and the field work continued until late May when caribou movement to calving

areas (located within the study area) is known to occur for some animals (Jenkins et al. in prep,

Elliot and Elliot 1975, 1974). Fortunately, to address the potential for delays, execution of the

survey was planned from south to north, in the direction of migration (Ferguson 1989, Redhead

and Land 1979, Elliot and Elliot 1975, Elliot and Elliot 1974). This ensured that calving areas

were surveyed towards the end of the field program and that any migrating animals were

counted. Finally, survey progress along the migration corridor was faster (ca. 40-50 km/day)

than the average daily movement rates of South Baffin caribou (average 2.26 km/day, SE 149 m;

derived from South Baffin caribou collaring data; May 1987-1991; Jenkins et al. in prep). We

acknowledge that long distance movements over short durations of time were possible, and

further analysis of available data is underway to expand our understanding. Notably, location

data for South Baffin is limited to 1987-1994, a period when caribou numbers were reported to

be much higher than our current estimate (Jenkins and Goorts 2013, Jenkins et al. 2012,

Ferguson and Gauthier 1992). In 2012, we detected very few animals along the migration

corridor suggesting that caribou are mainly non-migratory or too low in numbers, and that any

bias was negligible. These observations are supported by Inuit Knowledge shared by local

caribou experts Sandy Akavak and Kalola Pitsiulak, both from Kimmirut (Jenkins et al. 2012).

17

Sandy Akavak said “My father and the older ones used to say that the only spots where they

usually had good caribou were the ones that always had caribou. Our fathers used to say that

sometimes when the herd is declining the caribou stopped migrating, they stay in one spot. It’s

from the ancestors and my father. I have noticed in the last couple of years, in one area there’s

nothing but in other areas where there is always caribou, they are still there – it’s the only spot

and they stopped migrating (when the numbers are higher they migrate more).” Further, Kalola

Pitsiulak noted “When they decline, they seem to stop migrating. They seem to stay near the

calving grounds then.” While such information calls for more consultations to delineate these

specific areas, this suggests a possible density-dependent dispersion or scattering of animals,

where individuals in low density populations become even more isolated as their numbers

decline. Nevertheless, if caribou were possibly moving in the same direction that the study was

progressing, and moved faster than survey progress, a potential for positive bias is created by

double counting the animals.

Distance Sampling assumes that animals of interest are detected at their initial location

before responding in a non-random fashion to the observer (i.e. evasive movement away from

the transect). Such movement is likely negligible due to the slow speed of caribou relative to the

helicopter, but also because good detectability from the survey platform permitted early

sightings of animals. In the field we generally did not observe caribou running from the

helicopter except at distances very close to the transect center-line. In these cases, the open snow

covered environment captured obvious signs (e.g. fresh tracks) that were used as a marker for

location of first detection. Finally, inspection of the distribution of perpendicular distance data

did not detect any sign of violations of this assumption.

We followed Marques et al. 2006 to address measurement assumptions, specifically, that

the perpendicular distance (y) from the centerline to each detected cluster are measured

accurately along each transect line. Thus, we relied on post-sampling analyses using a GIS to

determine the perpendicular distance of clusters to a line given the plotted transect flown by the

helicopter pilot and overhead GPS position of caribou clusters at the point where clusters were

first observed. Using these protocols we suggest that measurement error was subsequently

negligible.

18

The most significant assumption of Distance Sampling is that animals that occur on the

transect are detected without exception, (i.e. g (0) = 1). The survey platform and winter schedule

were designed to address this assumption. We argue that caribou were unlikely to be missed

given that caribou are large, motile, animals in an open, treeless environment, and that snow

generally provided a white background and captured fresh caribou sign (e.g. tracks and feeding

sites). Notably, progress across the study area was from south to north (70° N), and snow

coverage persisted. Thus we suggest that our analysis, which assumes that caribou we passed

directly over were not missed, is reasonable. Importantly, even if g (0) was considerably less

than 1, the data and analyses still suggest that there are far fewer caribou in South Baffin than

previously estimated. For example even missing 50% of the caribou on the transect line (e.g.

because of weather or the ruggedness) would only double the abundance estimate to ca. 3000

animals.

The method of distance sampling was put at test here because of the huge area and the

low density population. To address heterogeneity, we used spatial stratification of the study area,

and the incorporation of covariates. Due in part, to the small number of observations in some

strata, the inclusion of habitat structure as a covariate in the analysis proved to be important. In

the future, additional variables, such as weather, snow cover, and observer, could also be

considered. Nonetheless, estimates provided in the present report are here as guidance for co-

management.

4.3 Consequences --.Overall, given the rigorous method, survey design and sampling effort, we

argue that the estimate is robust and our findings are consistent with information from local

communities (Jenkins et al. 2012, Jenkins and Goorts, 2013). Extensive consultations with

Hunter and Trapper Associations (HTOs) in Nov-Dec 2011, January-February 2012, and

December 2012 indicated that hunters have to travel further to locate caribou (Joannasie

Kooneeliusie, Elder, Qikiqtarjuaq, in Jenkins et al. 2012), and that there are few, if any caribou

in traditional harvesting areas (Joannasie Kooneeliusie, Qikiqtarjuaq; Sandy Akavak, Kimmirut;

Kamiata Nungusaigi, Cape Dorset, in Jenkins et al. 2012). Also, these consultations recorded

that declines in caribou have occurred over large areas, noting observations of dead caribou from

2008-2010 in relation to icing events (Joannasie Kooneeliusie, Qikiqtarjuaq, in Jenkins et al.

2012).

19

As no comparable surveys exist for the study area, we cannot determine the current trend

in abundance. However, Inuit Knowledge (IQ) indicates that caribou may have been declining

since the late 1990s and most significantly in the late 2000s (Jenkins et al. 2012). Based on

previous published estimates of population size (60,000-180,000; Ferguson and Gauthier 1992),

our results represent >95% decline in abundance. Thus, we recommend extensive consultations

with communities and/or community working groups to facilitate the co-development and

application of conservation and management measures. This will be the opportunity to gather

momentum into understanding what happened since the 1990s and how the dynamic of the

caribou on Baffin Island can be assessed and modelled. Inuit have traditional practices and laws

that may apply to situations of low abundance, and these should be considered when developing

the management framework (Sandy Akavak, Kimmirut; Kolola Pitsiulak, Kimmirut; in Jenkins

et al. 2012). This is a very challenging time for wildlife management where concrete efforts

among 10 communities and multiple stakeholders will be required to develop a meaningful and

timely plan to both conserve and manage these caribou.

20

5.0 REFERENCES:

Aars, J., T.A. Marques, S.T. Buckland, M. Andersen, S. Belikov, A. Boltunov, and O. Wiig.

2009. Estimating the Barents Sea Polar bear subpopulation size. Marine Mammal Science.

25(1): 35-52.

Buckland, S.T., D.R Anderson, K.P. Burnham, J.L. Laake, B.L. Borchers, and L. Thomas. 2004.

Advanced Distance Sampling. Estimating abundance of biological populations. Oxford

University Press, University of Oxford, Oxford, UK.

Buckland, S.T., K.P Burnham, J.L. Laake, D.L. Borchers, and L. Thomas. 2001. Introduction of

distance sampling: estimating abundance of biological populations. Oxford University Press,

New York.

Chowns, T. 1979. Wintering ground survey, South Baffin caribou herd, November 1979.

Renewable Resources unpublished report. Government of the Northwest Territories,

Yellowknife.

DoE, 2005. Recommendations on Total Allowable Harvest (TAH) Rates for Terrestrial Wildlife

Populations in Nunavut. Department of Environment. December 2005.

Ecological Stratification Working Group. 1995. A national ecological framework for Canada.

Agriculture and Agri-Food Canada, Research Branch, Center for Land and Biological Resources

Research and Environment Canada, State the Environment Directorate, Ecozone Analysis

Branch, Ottawa/Hull. Report and national map at 1:7 500 000 scale.

Elliot R.C. and C.E. Elliot. 1974. Baffin Island Caribou Surveys Progress Report, June 10-July

17, 1974. Game Management Service, Government of the Northwest Territories. Summary

Report.

Elliot R.C. and C.E. Elliot. 1975. Observation of the Distribution and Migration of Caribou on

Southern Baffin Island, July 4-August 1974. Game Management Service, government of the

Northwest Territories.

21

Ferguson, M.A.D. 1989. Baffin Island. In. Hall, E., ed. People and Caribou in the Northwest

Territories. Yellowknife; Renewable Resources, Government of the NWT. 141-149.

Ferguson M.A.D., and L. Gauthier. 1992. Status and trends of Rangifer tarandus and Ovibos

moschatus populations in Canada. Rangifer 12(3): 127-141.14

Ferguson, M.A.D. 2000. Mass Emigration of Arctic tundra caribou from a traditional winter

range: population dynamics and physical condition. Journal of Wildlife Management 64 (1);

168-178.

Ferguson, M.A.D. 2003. South Baffin caribou conservation and research plan: phase 1-

background and discussion report. Unpublished Report, Government of Nunavut, Department of

Environment, Pond Inlet, Nunavut, Canada.

Ferguson, M.A.D., Williams, R.G., and Messier, F. 1998. Inuit Knowledge of long-term changes

in a population of Arctic tundra caribou. Arctic 51: 201-219

Ferguson, R.S. 1995. Wildlife areas of special interest to the Department of Renewable

Resources I the Nunavut Settlement Area. Northwest Territories Department of Renewable

Resources, Wildlife Management Division. Yellowknife, NT.

Festa-Bianchet, M., J.C. Ray, S. Boutin, S.D. Coté, and A. Gunn. 2009. Conservation of caribou

(Rangifer tarandus) in Canada: an uncertain future. Canadian Journal of Zoology 89:419-434.

Hunters and Trappers Organizations of Broughton Island, Clyde River, Pond Inlet, Arctic Bay,

Igloolik, and Hall Beach, and the Qikiqtaaluk Wildlife Board. 1997. A Strategy for Future

Research on the North Baffin Caribou Population. Report. 13pp.

Jenkins, D. 2009. North Baffin Collaring Program. Field Summary Report. Department of

Environment, Government of Nunavut.

22

Jenkins, D. 2010. Space Use and Movement Pattersn of North Baffin Caribou. Interim Report,

January 2010. Department of Environment, Government of Nunavut.

Jenkins, D. 2011. Distribution and Abundance of Barrenground Caribou on Baffin Island,

Nunavut. Department of Environment, Government of Nunavut. Interim Report.

Jenkins, D. 2009a, 2010a, 2011a, and 2012. Distribution and Abundance of Barrenground

Caribou on Baffin Island, Nunavut. Research Proposal. Department of Environment,

Government of Nunavut.

Jenkins, D., and Goorts, J. 2011. Space use and movement patterns of North Baffin Caribou.

Field Summary and Progress Report, December 2011. Department of Environment, Government

of Nunavut.

Jenkins, D., Goorts, J., and Jeppessen, R. 2012. Baffin Island Caribou Consultations, 2012.

Draft.Department of Environment, Government of Nunavut.

Jenkins, D., and Goorts, J. 2013. Baffin Island Caribou Consultations, 2012. Draft. Department

of Environment, Government of Nunavut.

Jenkins et al. in prep. Movement and Space use by Barrenground Caribou on Baffin Island,

Nunavut: Canada’s Largest Island.

Jenkins, D.A., Campbell, M., Hope, G., Goorts, J., and McLoughlin, P. 2011. Recent trends in

abundance of Peary Caribou (Rangifer tarandus pearyi) and muskoxen (Ovibos moschatus) in

the Canadian Arctic Archipelago, Nunavut. Department of Environment, Government of

Nunavut, Wildlife Report No. 1, Pond Inlet, Nunavut. 184 pp.

Kelsall, J.P. 1949. Baffin Island Caribou Survey with Population Estimate and

Recommendations. Canadian Wildlife Service. Survey Report. 24 pp.

Kraft, Paul. 1984. Caribou Tagging on the Koukdjuak River, Baffin Island, N.W.T.: A Summary

and Analysis of Tag Returns. N.W.T. Wildlife Service. Progress Report No 10.

23

Marques, T.A., M. Andersen, S. Christensen-Dalsgaard, S. Belikov, A. Boltunov, O. Wiig, S.T,

Buckland, and J. Aars. 2006. The use of global positioning systems to record distances in a

helicopter line-transect survey. Wildlife Society Bulletin 34(4): 759-763.

Marques, T.A., L.Thomas, S.G. Fancy, and S.T. Buckland. 2007. Improving Estimates of Bird

Density Using Multiple Covariate Distance Sampling. Auk.

Milton Freeman Research Limited. 1976. Inuit land use and occupancy project. Volume 1: land

use and occupancy. Department of Indian and Northern Affairs. 23pp.

Redhead, R., and E. Land. 1979. Calving Ground Survey South Baffin Caribou Herd. June 1976.

Northwest Territories Wildlife Service, File Report – 2. 26 pp.

Rippin, Blair. 1972. Baffin Island Caribou. Game Management Division, Government of the

Northwest Territories, Survey Report.

Tener, J.S. Aerial Survey of Baffin Island Caribou, March 16-22, 1960. Canadian Wildlife

Service, Survey Report, 11 pp.

Thomas, L., Laake, J.L., Rexstad, E., Strindberg, S., Marques, F.F.C., Buckland, S.T., Borchers,

D.L., Anderson, D.R., Burnham, K.P., Burt, M.L., Hedley, S.L., Pollard, J.H., Bishop, J.R.B.

and Marques, T.A. 2009. Distance 6.0. Release 2. Research Unit for Wildlife Population

Assessment, University of St. Andrews, UK. http://www.ruwpa.st-and.ac.uk/distance/

Williams T.M., and Heard, D.C. 1986. World status of wild Rangifer tarandus populations.

Rangifer, special Issue No. 1: 19-28.

Wright, J.G. 1944. Economic Wildlife of Canada’s Eastern Arctic Caribou. Canadian

Geographic Journal, October: 184-19516

24

6.0 TABLES AND FIGURES

25

Figure 1a. Delineation of Barren-ground caribou populations on Baffin Island, as currently recognized (modified from DOE 2005).

Figure 1b. South Baffin Island caribou study area, 2012.

26

Figure 2. Ecoregions (as described by the Ecological Stratification Working Group, 1995) within the South Baffin study area (2012).

27

Table 1. Area (sq. m) surveyed in each ecoregion in the South Baffin survey area.

EcoRegion ID No. Area of Ecoregion in

Study Area (sq.m) Area of Ecoregion

Not surveyed (sq.m) Glaciated Area of Ecoregion (sq.m)

Area of Ecoregion Surveyed (sq.m)

Baffin Mountains 5 26012595670.674

4872533677.421 21140061993.253 Pangnirtung Upland 26 34270825269.184

608273899.173 33662551370.011

Hall Peninsula Upland 27 35388340567.403

126377399.698 35261963167.705 Meta Incognita Peninsula 28 71844146162.523

145402632.673 71698743529.850

Baffin Upland 29 16260276455.682

196814140.104 16063462315.577 Baffin Island Uplands 24 32202990217.572

18765207.435 32184225010.138

Foxe Basin Plain 25 44321511068.361 2059849192.963

42261661875.398 Melville Peninsula Plateau 23 26961031789.044

26961031789.044

Wager Bay Plateau 30 2452838712.381 2452838712.381

TOTAL 289714555912.824 4512687905.345 5968166956.505 279233701050.975 Note: The total area for each ecoregion was estimated using the North-Pole-Lambert-Azimuthal-Equa-Area projection centered on each ecoregion for independent area calculations. The ``Area of Ecoregion Not Surveyed` refers the portion of the ecoregion that was proposed to be surveyed in 2012, but was missed due to inclement weather and open water (see Figure 3b).

28

Figure 3a. South Baffin survey transect lines, spaced 10km apart and stratified by ecoregion.

Figure 3b. Combined survey track logs, recorded from March 27th to May 27th, 2012 by GPS units mounted in each helicopter. Three islands in Hudson Strait and 4 islands in Foxe Basin were not completed due to open water and inclement weather.

29

Table 2: Overall tally of observations and tracks recorded on and off- transect.

Species No. of

Observations Total No.

Individuals Comments No. of Tracks Caribou* 185 518

83

Polar Bear 19 36 20 Adults and 16 Cubs 36 Fox 13 14 11 Arctic, 1 Red, 2 not specified 51

Wolf 8 13

17 Arctic Hare 5 8

17

Bowhead Whale 4 15 Seal 1 2

Ptarmigan 10 31 Ducks (Eiders etc.) 8 1102+ 1089 Eider Ducks, 13 not specified

Geese 12 48+ 11 Snow, 34 Canada, 3 not specified Owl 5 5 4 Snowy, 1 not specified Birds 5 44+ 10 Snow Buntings, 34 not specified Unspecified

Tracks that were not caribou 87

Hunter/Skidoo 9 28 258+ *Caribou tracks were recorded as a priority; tracks for other wildlife species were not a priority and may be incomplete.

30

Table 3. Composition of caribou seen On and Off transect.

No. of Caribou Observed

Age Sex On Transect Off Transect Total

Adult

Male 94 35 129 Female 155 74 229 Unkmown 0 0 0 Total 249 109 358

Yearling

Male 42 11 53 Female 27 6 33 Unknown 3 4 7 Total 72 21 93

Calf

Male 11 0 11 Female 13 2 15 Unknown 10 14 24 Total 34 16 50

Unknown* Total 3 14 17

*Unknowns were not identified by age or sex. On-transect observations were included in the analysis as individuals could not be identified as calves.

31

Figure 4. Detection probability (continuous line) plot and histogram of perpendicular distances from the transect line for clusters of caribou in the South Baffin survey area, March 27-May27, 2012. The g(x) is estimated using a Half Normal model (see Table 5, below). Bin size is 165 m.

32

Table 4. Details of the MCDS and CDS Mdoels considered for the detection function, with available covariates. Covariates included course scale habitat structure (Coarse), and fine scale habitat structure (Fine). HN corresponds to half-normal and HR to hazard rate, C refers to Cosine and SP to Simple Polynomial. Columns are Par = number of parameters in the model, Delta AIC = the difference between lowest AIC (AKaike Information Criterion) and model AIC, D = global density estimate pooled across strata, D 95% CI = the corresponding 95% confidence interval for D, D CV = the coefficient of variation, as well as, N = estimated number of caribou in the survey area, N 95% CI = the corresponding 95% confidence interval for N, ESW = Effective Strip width, and three goodness of fit measures (GOF), Chi p = a chi-square test, K-S p = a Kolmogorov-Smirnov, and CvM = a Cramer-von Mises test.

Name Par Delta AIC D D 95% CI D CV N N 95% CI N CV ESW Goodness of Fit

AIC (km2) Lower Upper Lower Upper Chi p K-S p CvM HN_MCDS_Coarse 2 0.00 2118.171 0.0053 0.0038 0.0074 0.170 1484 1065 2067 0.17 1036.46 0.124 0.562 0.400 HN_MCDS _Fine 2 2.10 2120.275 0.0056 0.0040 0.0078 0.171 1552 1113 2165 0.17 1048.87 0.110 0.426 0.300 Hz_Sp_CDS* 2 15.94 2134.113 0.0064 0.0044 0.0093 0.191 1789 1234 2593 0.19 909.82 0.259 0.666 0.900 Hz_C_CDS* 2 15.94 2134.113 0.0064 0.0044 0.0093 0.193 1789 1230 2602 0.19 909.82 0.259 0.666 0.900 HN_C_CDS 2 17.37 2135.541 0.0059 0.0042 0.0084 0.177 1656 1174 2336 0.18 983.24 0.181 0.618 0.700 Hz_MCDS_Coarse 3 22.26 2140.431 0.0047 0.0031 0.0072 0.219 1314 859 2009 0.22 1238.77 0.040 0.049 0.050 HN_H_CDS* 1 24.25 2142.418 0.0048 0.0033 0.0070 0.195 1330 910 1943 0.20 1223.56 0.016 0.018 0.050 *Models run with warnings

33

Table 5: Estimates of caribou density and abundance within areas of eight different strata cover by line transect survey, based on analysis in the program Distance 6. The strata are nationally recognized ecoregions and BIU = Baffin Island Uplands, BU = Baffin Uplands, FBP = Foxe Basin Plain, HPU = Hall Peninsula Upland, MIP = Meta Incognita Peninsula, MPP= Melville Peninsula Plateau, PU= Pangnirtung Upland, BM=Baffin Mountains. Km Lines = the length of transect line surveyed in kilometers, Obs=number of distance observations used to fit the curves after truncation at 2800 m, D = estimated caribou density (per km2), N=estimated number of caribou in surveyed areas, CV = the coefficient of variation for D and N, and Km2

= the surveyed area.

Strata Km Lines Obs D D CV N N CV Km2 (km2) BIU 3280.00 3 0.00094 0.57 30 0.57 32184.23 BU 1595.42 3 0.00193 0.58 31 0.58 16063.46 FBP 4105.43 102 0.02556 0.21 1081 0.21 42261.66 HPU 3601.04 8 0.00228 0.43 81 0.43 35261.96 MIP 6968.16 8 0.00118 0.41 85 0.41 71698.74 MPP 2689.61 14 0.00535 0.40 144 0.40 26961.03 PU 3182.19 3 0.00097 0.58 33 0.58 33662.55 BM 1827.24 0 0 0.00 0 0.00 21140.06 Pooled 27249.88 141 0.00531 0.17 1484 0.17 279233.7